Identity theft and other related fraudulent identification activity has the potential to become a major problem to the economy, safety and stability of the United States. Identity theft refers to one individual fraudulently assuming the identity of another and may include activities such as opening credit cards in the name of another, obtaining loans, obtaining identification documents (e.g., drivers licenses, passports), obtaining entitlement/benefits cards (e.g., Social Security Cards, welfare cards, etc.), and the like. Often, these activities are performed without the consent or knowledge of the victim. Other fraudulent identification activity can also be problematic. An individual may, for example, use either his or her “real” identity to obtain a document, such as an identification card, but may further obtain additional identification cards using one or more identification credentials that belong to another and/or one or more fictitious identification credentials.
For example, to obtain an identification document such as a drivers license, a given individual may attempt to obtain multiple drivers licenses under different identities, may attempt to obtain a drivers license using false (e.g., “made up”), identification information, or may attempt to assume the identity of another to obtain a drivers license in that individual's name. In addition, individuals may alter legitimate identification documents to contain fraudulent information and may create wholly false identification documents that purport to be genuine documents.
It is extremely time consuming and expensive to apprehend and prosecute those responsible for identity theft and identity fraud. Thus, to help reduce identity theft and identity fraud, it may be advisable for issuers of identity-bearing documents to take affirmative preventative steps at the time of issuance of the identity documents. Because of the large number of documents that are issued every day and the large history of already issued documents, however, it is difficult for individual employees of the issuers to conduct effective searches at the time such documents are issued (or re-issued). In addition, the complexity and amount of the information stored often precludes manual searching, at least as a starting point.
For example, many government and business organizations, such as motor vehicle registries, store large databases of information about individuals. A motor vehicle registry database record may include information such as an operator's name, address, birth date, height, weight, and the like. Some motor vehicle registry databases also include images of the operator, such as a facial image and/or a fingerprint image. Unless the database is fairly small, it is nearly impossible for it to be searched manually.
In some databases, part or all of the database record is digitally encoded, which helps to make it possible to perform automated searches on the database. The databases themselves, however, can still be so large that automated searching is time consuming and error prone. For example, some states do not delete “old” images taken of a given individual. Each database record might be associated with a plurality of images. Thus, a database that contains records for 10 million individuals, could, in fact, contain 50-100 million images. If a given motor vehicle registry uses both facial and fingerprint images, the total number of images may be doubled still.
One promising search technique that can be used to perform automated searching of information and which may help to reduce identity theft and identity fraud is the use of biometric authentication and/or identification systems. Biometrics is a science that refers to technologies that can be used to measure and analyze physiological characteristics, such as eye retinas and irises, facial patterns, hand geometry, and fingerprints. Some biometrics technologies involve measurement and analysis of behavioral characteristics, such as voice patterns, signatures, and typing patterns. Because biometrics, especially physiological-based technologies, measures qualities that an individual usually cannot change, it can be especially effective for authentication and identification purposes.
Commercial manufacturers, such as Identix Corp of Minnetonka, Minn. manufacture biometric recognition systems that can be adapted to be capable of comparing two images. For example, the IDENTIX FACE IT product may be used to compare two facial images to determine whether the two images belong to the same person. Other commercial products are available that can compare two fingerprint images and determine whether the two images belong to the same person. For example, U.S. Pat. Nos. 6,072,894, 6,111,517, 6,185,316, 5,224,173, 5,450,504, and 5,991,429 further describe various types of biometrics systems, including facial recognition systems and fingerprint recognition systems, and these patents are hereby incorporated by reference in their entirety.
One difficulty in adapting commercial biometric systems to databases such as motor vehicle databases is the very large number of images that may be stored in the database. Some types of biometrics technologies can produce high numbers of false positives (falsely identifying a match between a first image and one or more other images) when the database size is very large. High numbers of false positives are sometimes seen with large databases of facial images that are used with facial recognition systems.
Another potential problem with searching large databases of biometric images can be the processing delays that can accompany so-called “one to many” searches (comparing a probe image with an “unidentified” image, such as a face or finger image presented for authentication, to a large database of previously enrolled “known” images. In addition, the “many” part of “one-to-many” can vary depending on the application and/or the biometric being used. In some types of applications (such as surveillance, terrorist watch lists, authentication for admission to a facility), the “many” can be as few as a few hundred individuals, whereas for other applications (e.g., issuance of security documents, such as passports, drivers licenses, etc.), the “many” can be many millions of images.
Further, some types of biometric technologies, such as facial recognition, have a few key differences from other types of biometric technologies, such as fingerprint technologies. For example, one difference between face recognition systems and fingerprint recognition systems can be cost. At the present time, for one to many type searching in identification document environments (where “many” at least means a million or more records), facial recognition systems are far less costly than fingerprint recognition systems. The more affordable fingerprint recognition systems, at the present time, generally include those adapted for one to few type searching (where “few” at least means fewer than a million records and includes, for example, systems adapted for use with tens of thousands of records).
Another difference between facial recognition systems and fingerprint recognition (and other systems, such as iris and retina identification systems, voice recognition systems, etc.) systems can be the error rates. In an exemplary biometric identification system, a given image (referred to as the “probe image”) is compared to one or more stored image to generate a candidate list of possible matches ordered by a match score. Like some types of automated fingerprint searches, at least some types of automated searches of facial images generate a candidate list of possible matches ordered by a match score. The score is a measure of the level of confidence that the probe facial image and a target image from the candidate list are portraits of the same person. With facial recognition systems, however, separating the true matches from the false matches can be much more difficult than with fingerprinting. Camera angles, angles at which the subject's head and/or eyes are turned, shadows, lighting, hats, glasses, beards, jewelry, etc., each have the potential to affect facial recognition results for at least some types of facial recognition systems. Thus, manual review of facial recognition results can be necessary.
Although face recognition systems can generate a relatively high percentage of matches in the candidate lists (with some facial recognition systems, the match percentage can be—90%), face recognition systems can also generate a very high number of false match results. Using a match threshold to define what constitutes a match may be less effective with face recognition than with other biometrics technologies (e.g., fingerprint recognition) because of the very high False Match Rates. Unless investigators are willing to manually verify matches in a face recognition candidate list, or a better technique is developed to differentiate between matches and non-matches, face recognition is likely to remain less effective than fingerprinting.
One measure of the accuracy of a given biometric systems is known as the Receiver Operating Curve (ROC). An ROC curve is a plot of a given systems False Match Rate (FMR) distribution against its False Non Match Rate (FNMR) distribution. Thus, ROC graphs for a given biometrics system show the relationship between the system's false match rate (a measure of the likelihood that the system will (incorrectly) match a subject with another, non-matching subject) and the false non-match rate (a measure of the likelihood that the system will fail to match a subject with another matching subject.)
In an office such as a Department of Motor Vehicles (DMV) office that utilizes one-to-many biometrics searching, the result of a false match for an applicant can be that the system identifies the person as matching the identity of another, different enrollee. The result of a false non-match for an applicant can be that the system fails to identify the person's additional, potentially fraudulent identities previously enrolled. Obviously, system implementers wish to minimize both false matches and false non-matches. However, ROC graphs may show that minimizing one problem tends to exacerbate the other. Tuning a biometrics system to identify the maximum number of possible fraudulent duplicate enrollees (minimizing false non-matches) may result in an increased number of enrollees being incorrectly identified as having multiple different identities (an increase of false matches). Conversely, a system tuned to minimize the number of applicants incorrectly identified as having multiple previously enrolled different identities will result in an increase in the number of possible fraudulent duplicate identities.
As noted previously, manual follow up searching is one way of augmenting biometric searching. A given biometrics search system can require verification of candidate lists because they will contain false matches, that is, subjects whom the system has falsely identified as matches. In some instances, the verification task is assigned to trained investigators who manually confirm each candidate's match. For example, the candidate list generated by a one-to-many fingerprint search would be manually checked by trained fingerprint analysts to eliminate the false matches contained in the list and verify the actual matches. Such follow up checking is time consuming and expensive.
We have discovered several techniques for augmenting and/or improving biometric search processes.
In a first aspect, we have found that a biometrics system that utilizes two different biometrics (a so-called “hybrid” biometrics system) can further automate and improve the biometrics search process, especially for (but not limited to) applications involving the production of identification documents such as drivers licenses. In one embodiment, the hybrid solution we propose combines the benefits of two biometric technologies: the speed of facial recognition and the accuracy of fingerprint matching. This blending of two biometrics can result in faster processing time and reduced costs. In one embodiment, this hybrid biometrics processing uses two different biometrics during the searching process, where the two different biometrics are used sequentially. The first biometric (e.g., a facial image) is used to retrieve a first set of results, and the first set of results are then searched using the second biometric (e.g., a fingerprint).
Depending on the particular pairing of biometric templates used, this process can provide a first biometric recognition process adapted for a one to many level “coarse” search to result in a candidate set of results that are usable by a second “one to few” type biometric recognition process. The set of results from the first biometric search process are therefore used by the second biometric search process. The inventors have found that use of two successive biometrics recognition processes (e.g., one to many facial for a first search, followed by one to (relatively) few fingerprint search) can significantly improve the accuracy of the overall recognition and may reduce the number of images that ultimately must be manually reviewed. This type of combination can result in higher accuracy searches at lower costs and/or in less time.
Selecting an appropriate combination also may result in higher accuracy searches at lower costs. For example, one to many fingerprint systems, at the present time, are so expensive that sometimes entities such as departments of Motor Vehicles (DMV's) cannot afford them, even though such one to many fingerprint searches often have desirable accuracy and may be easier to tune. In contrast, one to few fingerprint systems, at the present time, are significantly less expensive than one to many fingerprint systems, but are difficult to use with large databases. In contrast, one to many type facial recognition systems are relatively less expensive than one to many fingerprint systems, although they may be difficult to “tune” (as described above). In at least one embodiment of the invention, a system and method is provided that successively combines a one to many facial recognition system with a one to few fingerprint recognition system to achieve accuracy comparable to a one to many fingerprint system at about one third the cost of such a system.
This application describes additional inventive systems and methods for conducting biometrics searches and/or improving the accuracy of biometrics searching. We have found, for example, that using at least some of the returned biometric search results from a first initial probe image(whether or not the search is a hybrid type search) as second probe images can be used to “drill down” even further in the database and return even more high probability matches to the first probe image.
One embodiment provides a computerized system for determining whether a database contains an image substantially matching that of a given probe candidate. The system comprises an input device, a first database, a first biometric search engine, a second biometric search engine, and a processor. The input device is constructed and arranged to receive first and second biometric search templates associated with the probe candidate, the first biometric search template associated with a first type of biometric identifier and the second biometric search template associated with a second type of biometric identifier. The first database comprises a plurality of searchable biometric templates, the plurality of searchable biometric templates comprising a plurality of templates of the first type of biometric and a plurality of templates of the second type of biometric. The first biometric search engine is operably coupled to the input device and to the database and is adapted to search the database of searchable biometric templates for a match to the first biometric template and return a first set of results. The second biometric search engine is operably coupled to the input device and to the database and is adapted to search the database of searchable biometric templates for a match to the second biometric template and return a second set of results. The processor is in operable communication with the input device and the first and second biometric search engines. The processor programmed to compare the first and second sets of results to the first and second biometric templates to determine whether any result in the first set of results or the second set of results is a substantial match to either the first or second biometric search templates associated with the probe candidate.
Another embodiment provides a computer-implemented method for determining whether a database contains any images that substantially match at least one image provided of an individual. A probe data set is received, the comprising first and second biometric templates associated with the individual, the first biometric template associated with a different type of biometric than the second type of biometric template. A database of biometric templates is searched using the first biometric template to retrieve a first results set. A first predetermined portion of the first results set is selected. The first predetermined portion of the first results set is searched using the second biometric template to retrieve a second results set. A second predetermined portion of the second results set is selected. The second predetermined portion of the second results is provided for comparison with the image provided of the individual.
One aspect of the invention provides a computer-implemented method for determining whether a database contains any images that substantially match an image associated with an individual, comprising:
(a) receiving an initial probe data set, the first probe data set comprising at least one of a biometric template associated with the individual and non-biometric data associated with the individual;
(b) searching a database of biometric templates using the initial probe data set to retrieve a results set, the results set comprising biometric templates that satisfy a predetermined first criteria;
(c) selecting a first portion of the results set to be a refined probe data set;
(d) searching the database of biometric templates using the refined probe data set to return a results set, the results set comprising biometric templates that satisfy the predetermined first criteria;
(e) repeating (c) and (d) and returning a final results set.
Another aspect of the invention provides a computer-implemented method for determining whether a database contains any images that substantially match those of an individual, comprising:
(a) receiving an initial probe data set, the first probe data set comprising first and second biometric templates associated with the individual;
(b) searching a database of biometric templates using the first biometric template to retrieve a first results set, the first results set comprising a data set for each individual who has a biometric template that satisfies a predetermined first criteria, the data set comprising biometric template information that is searchable using the second biometric template;
(c) selecting a first portion of the first results set that satisfy a match criteria;
(d) searching the first portion of the first results set using the second biometric template to retrieve a second results set, the second results set comprising a data set for individuals who have a biometric template that satisfies a predetermined second criteria, the data set comprising biometric template information that is searchable using the first and second biometric templates;
(e) selecting a first portion of the second results set to be a refined probe data set; and
(f) searching the database of biometric templates using the refined probe data set to return a results set, the results set comprising biometric templates that satisfy a predetermined third criteria.
Another embodiment provides a method for locating images in a database, comprising:
receiving a first probe set, the first probe set comprising a non-biometric data record;
searching a database of data records for data records that substantially match the non biometric data record, the database of data records including, for each data record in the database, at least one biometric template associated with at least one image;
receiving a first results set from the search of the database, the results set comprising, for each substantial match to the non-biometric data record, a results set data record comprising a corresponding image and biometric template;
selecting at least one results set data record to use as a second probe set; and
searching the database of data records for data records having biometric templates that substantially match the biometric record associated with the results set data record.
In still another aspect, the invention provides a method for determining whether an individual should receive an identification document, comprising:
receiving a probe set associated with the individual, the probe set comprising a first biometric template of a first type and a second biometric template of a second type;
performing a first search, the first search comprising searching a database of previously enrolled biometric templates for a biometric template that substantially matches the biometric template of the first type;
performing a second search, the second search comprising searching the database of previously enrolled biometric templates for a biometric template that substantially matches the biometric template of the second type; and
analyzing the results of the first and second searches to determine whether any resulting matches indicate that the individual either is attempting to fraudulently obtain an identification document or has attempted to fraudulently obtain an identification document in the past.
The foregoing and other objects, aspects, features, and advantages of this invention will become even more apparent from the following description and drawings, and from the claims.
The drawings are not necessarily to scale, emphasis instead is generally placed upon illustrating the principles of the invention. In addition, in the drawings, like reference numbers indicate like elements. Further, in the Figures of this application, in some instances, a plurality of system elements or method steps may be shown as illustrative of a particular system element, and a single system element or method step may be shown as illustrative of a plurality of a particular systems elements or method steps. It should be understood that showing a plurality of a particular element or step is not intended to imply that a system or method implemented in accordance with the invention must comprise more than one of that element or step, nor is it intended by illustrating a single element or step that the invention is limited to embodiments having only a single one of that respective elements or steps. In addition, the total number of elements or steps shown for a particular system element or method is not intended to be limiting; those skilled in the art will recognize that the number of a particular system element or method steps can, in some instances, be selected to accommodate the particular user needs.